Research On Information Classification In The News Hall For Workshop

With the rapid development of network technologies,the Internet has become an important channel for people to obtain information.However,news information is chaotic and lacks correlation,so that people can not understand an event deeply and form a comprehensive and profound understanding.Classification is one of the key techniques to organize information.Traditional classification in the news field which just classifies news according to their categories,can not organize news information based on events effectively.Traditional sentiment classification analyses people's emotional tendencies,it can not mine event opinions in the round and solve the problem of people's one-sided understanding of events.Aiming at solving problems of traditional coarse classification and one-sided understanding of people,this dissertation puts forward a news hall for workshop,which draws on a principle of the meta-synthesis.The news hall organizes new information orderly based on events they describes,and extracts opinions of the events.According to the characteristics of the news hall,BFEDA(Event Detection Algorithm Based on Features)and BOICA(Information Classification Algorithm Based on Opinions of An Event)are proposed.The main work of the dissertation is reflected in the following aspects:(1)Focusing on solving the problem of the coarse classification,this dissertation proposes an algorithm named BFEDA.This algorithm extracts four features of an event from news to deal with the event.Aggregate similar news describing a same event qualitatively.And considering internal and external causes of events,events can be calculated quantitatively.Therefore,news information can be classified with the granularity of events and be organized effectively by this algorithm.(2)Concentrating on solving the problem of one-sided understanding of people,an algorithm named BOICA is proposed.BOICA extracts opinion sentences from web pages using words which can recognize opinion,and clusters the opinion sentences based on sentences similarity qualitatively.For each opinion cluster,BOICA gets the weight of opinions quantitatively by combining the credibility of web pages and the support degree of web page to opinion sentences.BOICA pays more attention to details of events and displays various views on events.(3)A prototype system of the news hall is designed and implemented.BFEDA algorithm and BOICA algorithm are experimented and analyzed by using the real data crawled from the Internet.The feasibility and effectiveness of BFEDA and BOICA are verified by the experimental results.So information can be classified in a fine-grained way and users can understand an event in a comprehensive and profound way.